Logistic regression: sigmoid and log loss

How logistic regression turns a linear score into a probability with the sigmoid, and fits it by maximum likelihood: minimising log loss (cross-entropy).

Part of Linear classifiers.

Read first

Read next

Related